Retinoblastoma is the most common intraocular tumor in childhood and is derived from the embryonal neuroectoderm. It shares some histopathological characteristic (partial or complete formation of rosettes) with other tumors of the central nervous system (i.e., neuroblastomas and medulloblastoma). In contrast to neuroblastomas, the cells of retinoblastomas are poorly characterized. A neuroepitheilial origin has been postulated for certain differentiating retinoblastomas, however, no biochemical markers have yet been described for these cells. The formation of Flexner-Wintersteiner rosettes and/or fleurettes, two morphological structures seen in some primary eye tumors, have been described as an imitation of the photoreceptor differentiation. We recently developed a new in vitro system which permits the routine growth and long term survival of all retinoblastomas obtained at surgery. The culture system also enables certain retinoblastomas to form well differentiated rosettes. It is the aim of the present proposal to use this system to study morphologically and biochemically human retinoblastoma cells cultured for a short time period in vitro. The process of photoreceptor cell differentiation can now be investigated as a function of time and changes in the ultrastructure (cell junctions, cilias, lamellar membrane stacks, neurosecretory granules) and the cell surface (microvillous projections) will be documented. Cells, actively participating in rosettes will be analyzed for DNA synthesis and cell division. Chemicals (cAMP, theophylline, Vitamin E) known to induce cellular differentiation with neuroblastoma cells will be tested for their abilities to induce photoreceptor cell differentiation. The presence of biochemical marker of retina cells (rhodopsin, catelcholamines, surface antigens) will be used for further characterization of retinoblastoma cells.